End-to-end ML pipeline for flood damage prediction using 30-day antecedent rainfall and catchment features. Benchmarks Linear Regression, Random Forest, XGBoost, SVR, and a Keras DNN with time-series-aware cross-validation and automated feature selection.

Actuarial study of water-supply interruptions, business losses, model validation and conditional provisions in Tunisia.

Machine Learning project that predicts Olympic medal counts using Linear Regression based on number of athletes and previous Olympic performance.

Can an ML model beat bookmaker odds? A rigorously validated 'no' — 65,900 ATP matches, XGBoost, chronological validation, 4 independent statistical tests.